diff options
Diffstat (limited to 'fs/ext4/migrate.c')
| -rw-r--r-- | fs/ext4/migrate.c | 338 |
1 files changed, 202 insertions, 136 deletions
diff --git a/fs/ext4/migrate.c b/fs/ext4/migrate.c index 5c1e27de775..ec092437d3e 100644 --- a/fs/ext4/migrate.c +++ b/fs/ext4/migrate.c @@ -12,21 +12,21 @@ * */ -#include <linux/module.h> -#include <linux/ext4_jbd2.h> -#include <linux/ext4_fs_extents.h> +#include <linux/slab.h> +#include "ext4_jbd2.h" +#include "ext4_extents.h" /* * The contiguous blocks details which can be * represented by a single extent */ -struct list_blocks_struct { - ext4_lblk_t first_block, last_block; +struct migrate_struct { + ext4_lblk_t first_block, last_block, curr_block; ext4_fsblk_t first_pblock, last_pblock; }; static int finish_range(handle_t *handle, struct inode *inode, - struct list_blocks_struct *lb) + struct migrate_struct *lb) { int retval = 0, needed; @@ -39,7 +39,7 @@ static int finish_range(handle_t *handle, struct inode *inode, newext.ee_block = cpu_to_le32(lb->first_block); newext.ee_len = cpu_to_le16(lb->last_block - lb->first_block + 1); ext4_ext_store_pblock(&newext, lb->first_pblock); - path = ext4_ext_find_extent(inode, lb->first_block, NULL); + path = ext4_ext_find_extent(inode, lb->first_block, NULL, 0); if (IS_ERR(path)) { retval = PTR_ERR(path); @@ -53,12 +53,14 @@ static int finish_range(handle_t *handle, struct inode *inode, * credit. But below we try to not accumalate too much * of them by restarting the journal. */ - needed = ext4_ext_calc_credits_for_insert(inode, path); + needed = ext4_ext_calc_credits_for_single_extent(inode, + lb->last_block - lb->first_block + 1, path); /* * Make sure the credit we accumalated is not really high */ - if (needed && handle->h_buffer_credits >= EXT4_RESERVE_TRANS_BLOCKS) { + if (needed && ext4_handle_has_enough_credits(handle, + EXT4_RESERVE_TRANS_BLOCKS)) { retval = ext4_journal_restart(handle, needed); if (retval) goto err_out; @@ -73,7 +75,7 @@ static int finish_range(handle_t *handle, struct inode *inode, goto err_out; } } - retval = ext4_ext_insert_extent(handle, inode, path, &newext); + retval = ext4_ext_insert_extent(handle, inode, path, &newext, 0); err_out: if (path) { ext4_ext_drop_refs(path); @@ -84,8 +86,7 @@ err_out: } static int update_extent_range(handle_t *handle, struct inode *inode, - ext4_fsblk_t pblock, ext4_lblk_t blk_num, - struct list_blocks_struct *lb) + ext4_fsblk_t pblock, struct migrate_struct *lb) { int retval; /* @@ -93,9 +94,10 @@ static int update_extent_range(handle_t *handle, struct inode *inode, */ if (lb->first_pblock && (lb->last_pblock+1 == pblock) && - (lb->last_block+1 == blk_num)) { + (lb->last_block+1 == lb->curr_block)) { lb->last_pblock = pblock; - lb->last_block = blk_num; + lb->last_block = lb->curr_block; + lb->curr_block++; return 0; } /* @@ -103,64 +105,49 @@ static int update_extent_range(handle_t *handle, struct inode *inode, */ retval = finish_range(handle, inode, lb); lb->first_pblock = lb->last_pblock = pblock; - lb->first_block = lb->last_block = blk_num; - + lb->first_block = lb->last_block = lb->curr_block; + lb->curr_block++; return retval; } static int update_ind_extent_range(handle_t *handle, struct inode *inode, - ext4_fsblk_t pblock, ext4_lblk_t *blk_nump, - struct list_blocks_struct *lb) + ext4_fsblk_t pblock, + struct migrate_struct *lb) { struct buffer_head *bh; __le32 *i_data; int i, retval = 0; - ext4_lblk_t blk_count = *blk_nump; unsigned long max_entries = inode->i_sb->s_blocksize >> 2; - if (!pblock) { - /* Only update the file block number */ - *blk_nump += max_entries; - return 0; - } - bh = sb_bread(inode->i_sb, pblock); if (!bh) return -EIO; i_data = (__le32 *)bh->b_data; - for (i = 0; i < max_entries; i++, blk_count++) { + for (i = 0; i < max_entries; i++) { if (i_data[i]) { retval = update_extent_range(handle, inode, - le32_to_cpu(i_data[i]), - blk_count, lb); + le32_to_cpu(i_data[i]), lb); if (retval) break; + } else { + lb->curr_block++; } } - - /* Update the file block number */ - *blk_nump = blk_count; put_bh(bh); return retval; } static int update_dind_extent_range(handle_t *handle, struct inode *inode, - ext4_fsblk_t pblock, ext4_lblk_t *blk_nump, - struct list_blocks_struct *lb) + ext4_fsblk_t pblock, + struct migrate_struct *lb) { struct buffer_head *bh; __le32 *i_data; int i, retval = 0; - ext4_lblk_t blk_count = *blk_nump; unsigned long max_entries = inode->i_sb->s_blocksize >> 2; - if (!pblock) { - /* Only update the file block number */ - *blk_nump += max_entries * max_entries; - return 0; - } bh = sb_bread(inode->i_sb, pblock); if (!bh) return -EIO; @@ -169,38 +156,28 @@ static int update_dind_extent_range(handle_t *handle, struct inode *inode, for (i = 0; i < max_entries; i++) { if (i_data[i]) { retval = update_ind_extent_range(handle, inode, - le32_to_cpu(i_data[i]), - &blk_count, lb); + le32_to_cpu(i_data[i]), lb); if (retval) break; } else { /* Only update the file block number */ - blk_count += max_entries; + lb->curr_block += max_entries; } } - - /* Update the file block number */ - *blk_nump = blk_count; put_bh(bh); return retval; } static int update_tind_extent_range(handle_t *handle, struct inode *inode, - ext4_fsblk_t pblock, ext4_lblk_t *blk_nump, - struct list_blocks_struct *lb) + ext4_fsblk_t pblock, + struct migrate_struct *lb) { struct buffer_head *bh; __le32 *i_data; int i, retval = 0; - ext4_lblk_t blk_count = *blk_nump; unsigned long max_entries = inode->i_sb->s_blocksize >> 2; - if (!pblock) { - /* Only update the file block number */ - *blk_nump += max_entries * max_entries * max_entries; - return 0; - } bh = sb_bread(inode->i_sb, pblock); if (!bh) return -EIO; @@ -209,16 +186,14 @@ static int update_tind_extent_range(handle_t *handle, struct inode *inode, for (i = 0; i < max_entries; i++) { if (i_data[i]) { retval = update_dind_extent_range(handle, inode, - le32_to_cpu(i_data[i]), - &blk_count, lb); + le32_to_cpu(i_data[i]), lb); if (retval) break; - } else + } else { /* Only update the file block number */ - blk_count += max_entries * max_entries; + lb->curr_block += max_entries * max_entries; + } } - /* Update the file block number */ - *blk_nump = blk_count; put_bh(bh); return retval; @@ -228,7 +203,7 @@ static int extend_credit_for_blkdel(handle_t *handle, struct inode *inode) { int retval = 0, needed; - if (handle->h_buffer_credits > EXT4_RESERVE_TRANS_BLOCKS) + if (ext4_handle_has_enough_credits(handle, EXT4_RESERVE_TRANS_BLOCKS+1)) return 0; /* * We are freeing a blocks. During this we touch @@ -236,7 +211,7 @@ static int extend_credit_for_blkdel(handle_t *handle, struct inode *inode) * So allocate a credit of 3. We may update * quota (user and group). */ - needed = 3 + 2*EXT4_QUOTA_TRANS_BLOCKS(inode->i_sb); + needed = 3 + EXT4_MAXQUOTAS_TRANS_BLOCKS(inode->i_sb); if (ext4_journal_extend(handle, needed) != 0) retval = ext4_journal_restart(handle, needed); @@ -260,13 +235,17 @@ static int free_dind_blocks(handle_t *handle, for (i = 0; i < max_entries; i++) { if (tmp_idata[i]) { extend_credit_for_blkdel(handle, inode); - ext4_free_blocks(handle, inode, - le32_to_cpu(tmp_idata[i]), 1, 1); + ext4_free_blocks(handle, inode, NULL, + le32_to_cpu(tmp_idata[i]), 1, + EXT4_FREE_BLOCKS_METADATA | + EXT4_FREE_BLOCKS_FORGET); } } put_bh(bh); extend_credit_for_blkdel(handle, inode); - ext4_free_blocks(handle, inode, le32_to_cpu(i_data), 1, 1); + ext4_free_blocks(handle, inode, NULL, le32_to_cpu(i_data), 1, + EXT4_FREE_BLOCKS_METADATA | + EXT4_FREE_BLOCKS_FORGET); return 0; } @@ -295,7 +274,9 @@ static int free_tind_blocks(handle_t *handle, } put_bh(bh); extend_credit_for_blkdel(handle, inode); - ext4_free_blocks(handle, inode, le32_to_cpu(i_data), 1, 1); + ext4_free_blocks(handle, inode, NULL, le32_to_cpu(i_data), 1, + EXT4_FREE_BLOCKS_METADATA | + EXT4_FREE_BLOCKS_FORGET); return 0; } @@ -306,8 +287,10 @@ static int free_ind_block(handle_t *handle, struct inode *inode, __le32 *i_data) /* ei->i_data[EXT4_IND_BLOCK] */ if (i_data[0]) { extend_credit_for_blkdel(handle, inode); - ext4_free_blocks(handle, inode, - le32_to_cpu(i_data[0]), 1, 1); + ext4_free_blocks(handle, inode, NULL, + le32_to_cpu(i_data[0]), 1, + EXT4_FREE_BLOCKS_METADATA | + EXT4_FREE_BLOCKS_FORGET); } /* ei->i_data[EXT4_DIND_BLOCK] */ @@ -327,7 +310,7 @@ static int free_ind_block(handle_t *handle, struct inode *inode, __le32 *i_data) } static int ext4_ext_swap_inode_data(handle_t *handle, struct inode *inode, - struct inode *tmp_inode) + struct inode *tmp_inode) { int retval; __le32 i_data[3]; @@ -339,7 +322,7 @@ static int ext4_ext_swap_inode_data(handle_t *handle, struct inode *inode, * i_data field of the original inode */ retval = ext4_journal_extend(handle, 1); - if (retval != 0) { + if (retval) { retval = ext4_journal_restart(handle, 1); if (retval) goto err_out; @@ -351,10 +334,21 @@ static int ext4_ext_swap_inode_data(handle_t *handle, struct inode *inode, down_write(&EXT4_I(inode)->i_data_sem); /* + * if EXT4_STATE_EXT_MIGRATE is cleared a block allocation + * happened after we started the migrate. We need to + * fail the migrate + */ + if (!ext4_test_inode_state(inode, EXT4_STATE_EXT_MIGRATE)) { + retval = -EAGAIN; + up_write(&EXT4_I(inode)->i_data_sem); + goto err_out; + } else + ext4_clear_inode_state(inode, EXT4_STATE_EXT_MIGRATE); + /* * We have the extent map build with the tmp inode. * Now copy the i_data across */ - ei->i_flags |= EXT4_EXTENTS_FL; + ext4_set_inode_flag(inode, EXT4_INODE_EXTENTS); memcpy(ei->i_data, tmp_ei->i_data, sizeof(ei->i_data)); /* @@ -390,7 +384,7 @@ static int free_ext_idx(handle_t *handle, struct inode *inode, struct buffer_head *bh; struct ext4_extent_header *eh; - block = idx_pblock(ix); + block = ext4_idx_pblock(ix); bh = sb_bread(inode->i_sb, block); if (!bh) return -EIO; @@ -406,7 +400,8 @@ static int free_ext_idx(handle_t *handle, struct inode *inode, } put_bh(bh); extend_credit_for_blkdel(handle, inode); - ext4_free_blocks(handle, inode, block, 1, 1); + ext4_free_blocks(handle, inode, NULL, block, 1, + EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET); return retval; } @@ -431,28 +426,27 @@ static int free_ext_block(handle_t *handle, struct inode *inode) return retval; } return retval; - } -int ext4_ext_migrate(struct inode *inode, struct file *filp, - unsigned int cmd, unsigned long arg) +int ext4_ext_migrate(struct inode *inode) { handle_t *handle; int retval = 0, i; __le32 *i_data; - ext4_lblk_t blk_count = 0; struct ext4_inode_info *ei; struct inode *tmp_inode = NULL; - struct list_blocks_struct lb; + struct migrate_struct lb; unsigned long max_entries; + __u32 goal; + uid_t owner[2]; - if (!test_opt(inode->i_sb, EXTENTS)) - /* - * if mounted with noextents we don't allow the migrate - */ - return -EINVAL; - - if ((EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL)) + /* + * If the filesystem does not support extents, or the inode + * already is extent-based, error out. + */ + if (!EXT4_HAS_INCOMPAT_FEATURE(inode->i_sb, + EXT4_FEATURE_INCOMPAT_EXTENTS) || + (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) return -EINVAL; if (S_ISLNK(inode->i_mode) && inode->i_blocks == 0) @@ -461,34 +455,35 @@ int ext4_ext_migrate(struct inode *inode, struct file *filp, */ return retval; - handle = ext4_journal_start(inode, - EXT4_DATA_TRANS_BLOCKS(inode->i_sb) + - EXT4_INDEX_EXTRA_TRANS_BLOCKS + 3 + - 2 * EXT4_QUOTA_INIT_BLOCKS(inode->i_sb) - + 1); + /* + * Worst case we can touch the allocation bitmaps, a bgd + * block, and a block to link in the orphan list. We do need + * need to worry about credits for modifying the quota inode. + */ + handle = ext4_journal_start(inode, EXT4_HT_MIGRATE, + 4 + EXT4_MAXQUOTAS_TRANS_BLOCKS(inode->i_sb)); + if (IS_ERR(handle)) { retval = PTR_ERR(handle); - goto err_out; + return retval; } - tmp_inode = ext4_new_inode(handle, - inode->i_sb->s_root->d_inode, - S_IFREG); + goal = (((inode->i_ino - 1) / EXT4_INODES_PER_GROUP(inode->i_sb)) * + EXT4_INODES_PER_GROUP(inode->i_sb)) + 1; + owner[0] = i_uid_read(inode); + owner[1] = i_gid_read(inode); + tmp_inode = ext4_new_inode(handle, inode->i_sb->s_root->d_inode, + S_IFREG, NULL, goal, owner); if (IS_ERR(tmp_inode)) { - retval = -ENOMEM; + retval = PTR_ERR(tmp_inode); ext4_journal_stop(handle); - tmp_inode = NULL; - goto err_out; + return retval; } i_size_write(tmp_inode, i_size_read(inode)); /* - * We don't want the inode to be reclaimed - * if we got interrupted in between. We have - * this tmp inode carrying reference to the - * data blocks of the original file. We set - * the i_nlink to zero at the last stage after - * switching the original file to extent format + * Set the i_nlink to zero so it will be deleted later + * when we drop inode reference. */ - tmp_inode->i_nlink = 1; + clear_nlink(tmp_inode); ext4_ext_tree_init(handle, tmp_inode); ext4_orphan_add(handle, tmp_inode); @@ -498,17 +493,33 @@ int ext4_ext_migrate(struct inode *inode, struct file *filp, * start with one credit accounted for * superblock modification. * - * For the tmp_inode we already have commited the - * trascation that created the inode. Later as and + * For the tmp_inode we already have committed the + * transaction that created the inode. Later as and * when we add extents we extent the journal */ /* - * inode_mutex prevent write and truncate on the file. Read still goes - * through. We take i_data_sem in ext4_ext_swap_inode_data before we - * switch the inode format to prevent read. + * Even though we take i_mutex we can still cause block + * allocation via mmap write to holes. If we have allocated + * new blocks we fail migrate. New block allocation will + * clear EXT4_STATE_EXT_MIGRATE flag. The flag is updated + * with i_data_sem held to prevent racing with block + * allocation. */ - mutex_lock(&(inode->i_mutex)); - handle = ext4_journal_start(inode, 1); + down_read(&EXT4_I(inode)->i_data_sem); + ext4_set_inode_state(inode, EXT4_STATE_EXT_MIGRATE); + up_read((&EXT4_I(inode)->i_data_sem)); + + handle = ext4_journal_start(inode, EXT4_HT_MIGRATE, 1); + if (IS_ERR(handle)) { + /* + * It is impossible to update on-disk structures without + * a handle, so just rollback in-core changes and live other + * work to orphan_list_cleanup() + */ + ext4_orphan_del(NULL, tmp_inode); + retval = PTR_ERR(handle); + goto out; + } ei = EXT4_I(inode); i_data = ei->i_data; @@ -516,35 +527,32 @@ int ext4_ext_migrate(struct inode *inode, struct file *filp, /* 32 bit block address 4 bytes */ max_entries = inode->i_sb->s_blocksize >> 2; - for (i = 0; i < EXT4_NDIR_BLOCKS; i++, blk_count++) { + for (i = 0; i < EXT4_NDIR_BLOCKS; i++) { if (i_data[i]) { retval = update_extent_range(handle, tmp_inode, - le32_to_cpu(i_data[i]), - blk_count, &lb); + le32_to_cpu(i_data[i]), &lb); if (retval) goto err_out; - } + } else + lb.curr_block++; } if (i_data[EXT4_IND_BLOCK]) { retval = update_ind_extent_range(handle, tmp_inode, - le32_to_cpu(i_data[EXT4_IND_BLOCK]), - &blk_count, &lb); + le32_to_cpu(i_data[EXT4_IND_BLOCK]), &lb); if (retval) goto err_out; } else - blk_count += max_entries; + lb.curr_block += max_entries; if (i_data[EXT4_DIND_BLOCK]) { retval = update_dind_extent_range(handle, tmp_inode, - le32_to_cpu(i_data[EXT4_DIND_BLOCK]), - &blk_count, &lb); + le32_to_cpu(i_data[EXT4_DIND_BLOCK]), &lb); if (retval) goto err_out; } else - blk_count += max_entries * max_entries; + lb.curr_block += max_entries * max_entries; if (i_data[EXT4_TIND_BLOCK]) { retval = update_tind_extent_range(handle, tmp_inode, - le32_to_cpu(i_data[EXT4_TIND_BLOCK]), - &blk_count, &lb); + le32_to_cpu(i_data[EXT4_TIND_BLOCK]), &lb); if (retval) goto err_out; } @@ -559,9 +567,15 @@ err_out: * tmp_inode */ free_ext_block(handle, tmp_inode); - else - retval = ext4_ext_swap_inode_data(handle, inode, - tmp_inode); + else { + retval = ext4_ext_swap_inode_data(handle, inode, tmp_inode); + if (retval) + /* + * if we fail to swap inode data free the extent + * details of the tmp inode + */ + free_ext_block(handle, tmp_inode); + } /* We mark the tmp_inode dirty via ext4_ext_tree_init. */ if (ext4_journal_extend(handle, 1) != 0) @@ -584,19 +598,71 @@ err_out: /* Reset the extent details */ ext4_ext_tree_init(handle, tmp_inode); + ext4_journal_stop(handle); +out: + unlock_new_inode(tmp_inode); + iput(tmp_inode); - /* - * Set the i_nlink to zero so that - * generic_drop_inode really deletes the - * inode - */ - tmp_inode->i_nlink = 0; + return retval; +} - ext4_journal_stop(handle); - mutex_unlock(&(inode->i_mutex)); +/* + * Migrate a simple extent-based inode to use the i_blocks[] array + */ +int ext4_ind_migrate(struct inode *inode) +{ + struct ext4_extent_header *eh; + struct ext4_super_block *es = EXT4_SB(inode->i_sb)->s_es; + struct ext4_inode_info *ei = EXT4_I(inode); + struct ext4_extent *ex; + unsigned int i, len; + ext4_fsblk_t blk; + handle_t *handle; + int ret; + + if (!EXT4_HAS_INCOMPAT_FEATURE(inode->i_sb, + EXT4_FEATURE_INCOMPAT_EXTENTS) || + (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) + return -EINVAL; - if (tmp_inode) - iput(tmp_inode); + if (EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb, + EXT4_FEATURE_RO_COMPAT_BIGALLOC)) + return -EOPNOTSUPP; - return retval; + handle = ext4_journal_start(inode, EXT4_HT_MIGRATE, 1); + if (IS_ERR(handle)) + return PTR_ERR(handle); + + down_write(&EXT4_I(inode)->i_data_sem); + ret = ext4_ext_check_inode(inode); + if (ret) + goto errout; + + eh = ext_inode_hdr(inode); + ex = EXT_FIRST_EXTENT(eh); + if (ext4_blocks_count(es) > EXT4_MAX_BLOCK_FILE_PHYS || + eh->eh_depth != 0 || le16_to_cpu(eh->eh_entries) > 1) { + ret = -EOPNOTSUPP; + goto errout; + } + if (eh->eh_entries == 0) + blk = len = 0; + else { + len = le16_to_cpu(ex->ee_len); + blk = ext4_ext_pblock(ex); + if (len > EXT4_NDIR_BLOCKS) { + ret = -EOPNOTSUPP; + goto errout; + } + } + + ext4_clear_inode_flag(inode, EXT4_INODE_EXTENTS); + memset(ei->i_data, 0, sizeof(ei->i_data)); + for (i=0; i < len; i++) + ei->i_data[i] = cpu_to_le32(blk++); + ext4_mark_inode_dirty(handle, inode); +errout: + ext4_journal_stop(handle); + up_write(&EXT4_I(inode)->i_data_sem); + return ret; } |
